CN104639472B - Uplink multi-users mimo channel method of estimation - Google Patents

Abstract

The present application discloses an uplink multi-user MIMO channel estimation method, which includes: the base station utilizes the pilot sequence S ₀ (k,l) of a MU-MIMO user UE ₀ for resource multiplexing transmission in the uplink, and calculates the received MU-MIMO The MIMO reference signal Y(k,l) performs least square LS channel estimation to obtain the channel estimation result Wherein, k is the carrier number, k=1,2,...,K, K is the number of subcarriers per resource block, and l is the number of OFDM pilot symbols currently performing channel estimation; Equally divided into K/N segments and perform N-order discrete Fourier transform DFT transform to obtain the LS channel estimation sequence of each MU-MIMO user performing resource multiplexing transmission, where N is the number of MU-MIMO users performing resource multiplexing transmission number; for the LS channel estimation sequence of each MU-MIMO user performing resource multiplexing transmission, respectively perform minimum mean square error criterion MMSE filtering to obtain the channel estimation result of each MU-MIMO user. The present invention can effectively reduce the complexity of channel estimation.

Description

Uplink multi-user MIMO channel estimation method

technical field

The invention relates to mobile communication technology, in particular to an uplink multi-user multiple-input multiple-output (MIMO) channel estimation method in an LTE system.

Background technique

In order to solve the contradiction between limited wireless resources and increasing mass data transmission, wireless communication systems must have higher communication capacity and higher spectral efficiency. As the core technology of LTE communication system, MIMO technology has become an inevitable choice to realize the use of limited space resources to improve spectrum efficiency. Uplink 2-user MU-MIMO is already one of the necessary functional requirements of the existing LTE public network system. In order to cope with the foreseeable further growth of user demand, uplink MU-MIMO with 3 users or even more users can increase the system capacity more. Although MIMO technology utilizes the wireless channel feature independence between multiple antennas to improve channel capacity, the reception of MIMO signals becomes more complicated with the increase of simultaneous frequency code streams, which leads to conflicts between MIMO processing and LTE system requirements for real-time performance and low power consumption. exacerbated. Channel estimation is a crucial part of signal reception and processing, and the balance of its algorithm performance and complexity is of great significance to correct signal reception and real-time operation of the system.

The uplink service transmission of the LTE system all carries the Zadoff-Chu sequence pilot signal, so the channel estimation of the receiver uses the pilot-assisted channel estimation. Common channel estimates include least square (LS) and minimum mean square error criterion (MMSE) channel estimation, as well as LS estimation for time domain denoising, linear filtering MMSE estimation and various other extensions.

The basic principle of LS channel estimation is to perform channel estimation based on the principle of the minimum square of the difference between the received pilot and the known pilot, that is, H _LS =agrmin|Y-SH _LS |, where H _LS is the LS channel estimation result on the carrier, Y is the received pilot value and S is the known transmitted pilot value. LS channel estimation with time domain denoising is a channel estimation method that converts the LS channel estimation sequence results to the time domain through IDFT, denoises and filters the time domain data, and then transforms the denoised time domain sequence through DFT to obtain the channel estimation results in the frequency domain. . The LS channel estimation method does not use the correlation of the channel transfer function of the subcarriers, nor does it consider the influence of noise, and the estimation error caused by the noise cannot be eliminated; the performance of time-domain denoising LS channel estimation and other evolution methods is slightly better than that of LS estimation. However, due to the multiplied number of IDFT/DFT in the MIMO scenario, its implementation procedure is more complicated. In addition, the correlation of each point of the transfer function is not considered, and the channel estimation performance cannot meet the MU-MIMO application.

Theoretically, the MMSE channel estimation algorithm needs to perform two-dimensional processing in the time domain and frequency domain. After initial simplification, filtering in the frequency domain can be expressed as H _MMSE = R _HH (R _HH +σ ² (SS ^H ) ^-1 ) ^-1 H _LS , where H _MMSE is the MMSE channel estimation result, R _HH is the channel autocorrelation matrix, and ^σ2 is the additive white Gaussian noise (AGWN). The MMSE algorithm has good channel estimation performance, but it is very complicated to calculate the channel autocorrelation matrix in real time. In LTE system applications, the commonly used MMSE channel estimation is a more simplified algorithm: H _MMSE = WH _LS , W is a filter pre-designed according to the AGWN channel parameters, that is, linear MMSE filtering is performed on the LS channel estimation results, which can be called the The method is a simplified LMMSE channel estimation method. The simplified LMMSE channel estimation method is relatively simple to implement, and its algorithm performance is superior to that of LS channel estimation and time-domain denoising LS channel estimation. Universally applicable channel estimation method. However, since the LMMSE channel estimation method is to filter the LS channel estimation results of all users, when the number of users is large, the order of the filter will increase, so the complexity of the filter will increase, while in MU-MIMO In the scenario, the number of users carried on the same time-frequency resource is doubled. Therefore, the LMMSE channel estimation processing time and resources also need to be doubled, which affects the real-time performance and power consumption of the system.

Other LS estimation and MMSE estimation extension methods, or frequency offset compensation based on LS estimation, or special design of MMSE filter coefficients, compared with LMMSE channel estimation, channel estimation has higher complexity and performance improvement is small or Even worse, it is not conducive to the application of MU-MIMO systems.

To sum up, the above-mentioned existing channel estimation methods have a problem of high complexity when applied to MU-MIMO scenarios, and are not suitable for MU-MIMO systems.

Contents of the invention

In view of this, the main purpose of the present invention is to provide an uplink multi-user MIMO channel estimation method, which can effectively reduce the complexity of channel estimation.

In order to achieve the above object, the technical scheme proposed by the present invention is:

A method for uplink multi-user MIMO channel estimation, comprising:

a. The base station uses the pilot sequence S ₀ (k,l) of an MU-MIMO user UE ₀ for uplink resource multiplexing transmission, and performs the least square LS channel on the received MU-MIMO reference signal Y(k,l). Estimated to get the channel estimation result ; Wherein, k is the carrier number, k=1,2,...,K, K is the number of subcarriers per resource block, and l is the number of OFDM pilot symbols currently performing channel estimation;

b. The K described Equally divided into K/N segments and perform N-order discrete Fourier transform DFT transform to obtain the LS channel estimation sequence of each MU-MIMO user performing resource multiplexing transmission, where N is the number of MU-MIMO users performing resource multiplexing transmission number;

c. For the LS channel estimation sequence of each MU-MIMO user performing resource multiplexing transmission, perform minimum mean square error criterion MMSE filtering respectively to obtain a channel estimation result of each MU-MIMO user.

In summary, the uplink multi-user MIMO channel estimation method proposed by the present invention first determines the LS channel estimation results of each user, and then performs MMSE filtering on the LS channel estimation results of each user, so that the filter can be effectively simplified. complexity, which in turn can effectively reduce the complexity of channel estimation.

Description of drawings

Fig. 1 is a schematic flow chart of Embodiment 1 of the present invention;

Figure 2 is a curve diagram of user 0 signal-to-noise ratio --- block error rate under the SCME channel of the micro cell;

Figure 3 is a curve diagram of user 1 signal-to-noise ratio --- block error rate under the SCME channel of the micro cell;

Fig. 4 is a curve diagram of user 2 signal-to-noise ratio --- block error rate under the SCME channel of the micro cell;

Fig. 5 is a user 0 signal-to-noise ratio --- block error rate curve diagram under the SCME channel of the macro cell;

Fig. 6 is a user 1 signal-to-noise ratio --- block error rate curve diagram under the SCME channel of the macro cell;

Fig. 7 is a curve diagram of user 2 signal-to-noise ratio-block error rate under the SCME channel of the macro cell.

detailed description

In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

The core idea of the present invention is: firstly determine the LS channel estimation results of each user, and then perform MMSE filtering on the LS channel estimation results of each user respectively. In this way, since the filter only needs to filter the LS channel estimation results of a single user , therefore, the complexity of the filter can be effectively simplified and the complexity of channel estimation can be reduced.

For ease of understanding, before the specific embodiments of the present invention are described, the implementation principle of how to obtain the LS channel estimation results of each user in the present invention is described as follows:

The pilot of the LTE uplink traffic channel is a Zadoff-Chu sequence. According to the LTE protocol (3GPP TS36.211), the pilot sequences of N MU users will have the following relationship: Where S _n+1 (k,l) and S _n (k,l) are the pilots of user n+1 and user n at carrier k of symbol l respectively, and delta_n _cs is an integer in the range of [1,11]. Considering the minimization of mutual interference after MU user pilot superposition, the number of MU paired users N is limited to an integer factor of 12 (such as 2, 3, 4, 6), and at the same time, delta_n _cs is set to 12/N, then at this time

From the above analysis, it can be seen that the N subcarrier points of the reference signal sequence of user n and user 0 (the sequence length is K, and K is an integer multiple of 12) have the following relationship:

At subcarrier Nm+i Among them, i=0,1,...,N-1, m=0,1,2,..., K/N-1, N users of MU-MIMO actually form an N-order DFT code division multiplexing relationship, User 0 is any one of the N MU users, and the numbers of the other N-1 users can be determined according to the phase relationship with user 0, and the specific method is the same as that of the existing system.

The N-user wireless signals of MU-MIMO are superimposed during free space transmission. After passing through the AWGN channel, the pilot signal Y(k,l) of subcarrier k and symbol l received by the receiving end can be expressed as follows:

Y(k,l)=H ₀ (k,l)S ₀ (k,l)+H ₁ (k,l)S ₁ (k,l)+…+H _N-1 (k,l)S _{N -1} (k,l)+Noise(k,l)

Among them, H and S represent the channel response and the transmission pilot respectively, the subscript is the user ID, and Noise represents the AWGN channel noise.

Preliminary LS channel estimation is performed on the received MU-MIMO reference signal with the local pilot sequence S ₀ (k,l) of user 0:

Wherein, S ^* ₀ (k,l) is the conjugate of S ₀ (k,l).

Using the relationship between the N subcarriers in the local reference sequences S _n (k,l) and S ₀ (k,l) of user n and user 0 will be of length K sequence (ie ) are equally divided into K/N segments, which can be deduced Among the N subcarriers of each segment of , the LS channel estimation of subcarrier i is as follows:

In general, if the signal coherence bandwidth is greater than 1 PRB, then the channel response H(k,l) of each user subcarrier within the coherence bandwidth is almost unchanged, and the ideal channel estimate H _n (Nm +i,l) can be replaced by H _n (Nm,l), that is, the above equation can be approximately written as:

Formula 1

by is an independent sequence X, H ₀ (Nm,l), H ₁ (Nm,l), ..., H _N-1 (Nm,l) is an independent sequence Z, as can be seen from the above derivation, X and There is an IDFT relationship between Z, namely:

N′(i) is Noise(Nm+i,l)S ^* ₀ (Nm+i,l) Formula 2

From this, the sequence Z can be solved by N-order DFT:

Formula 3

According to Equation 3, the channel estimation of user n can be obtained from the channel estimation of user 0 through the following formula:

Formula 4

Since Noise(Nm+i,l)S ^* ₀ (Nm+i,l) is unknown when estimating the channel estimate of user n, only the LS channel estimate of user n can be obtained Expressed as:

Formula 5

In this way, according to formula 5, the LS channel estimation value of each user can be obtained, that is, the LS channel estimation value of length K The sequence is equally divided into M=K/N segments for N-order DFT transformation to obtain LS channel estimates for N users.

FIG. 1 is a schematic flow chart of Embodiment 1 of the present invention. As shown in FIG. 1 , the uplink multi-user MIMO channel estimation method of this embodiment mainly includes:

Step 101, the base station uses the pilot sequence S ₀ (k,l) of a MU-MIMO user UE0 for uplink resource multiplexing transmission, and performs the least square LS channel on the received MU-MIMO reference signal Y(k,l) Estimated to get the channel estimation result

Wherein, k is the carrier number, k=1,2,...,K, K is the number of subcarriers per resource block, and l is the number of the OFDM pilot symbol currently performing channel estimation.

Like the existing system, the resource block is composed of K subcarrier resources of one OFDM symbol, and the channel estimation processing unit of the present invention is one OFDM symbol where the pilot symbol is located, that is, one OFDM pilot symbol.

Specifically, in this step, follow the The specific method for performing the LS channel estimation is known by those skilled in the art, and will not be repeated here.

Here, UE ₀ is any MU-MIMO user performing uplink resource multiplexing transmission. Same as the existing system, after UE ₀ is determined, the numbers of other users can be determined according to the phase relationship.

Step 102, the K described Divide into K/N segments and perform N-order discrete Fourier transform (DFT) to obtain the LS channel estimation sequence of each MU-MIMO user for resource multiplexing transmission.

Wherein, N is the number of MU-MIMO users performing resource multiplexing transmission.

This step is used to determine the LS channel estimation sequence of each user, so as to reduce the complexity of subsequent filtering and denoising, thereby effectively reducing the complexity of N-user MU-MIMO channel estimation.

Specifically, as mentioned above, this step can be realized by the following methods:

use according to Obtain the sequence of least squares (LS) channel estimates for each of the MU-MIMO user n

Among them, m=0,1,2,..., K/N-1, n is the user number, n=0,1,...,N-1.

Step 103: For the LS channel estimation sequence of each MU-MIMO user performing resource multiplexing transmission, respectively perform minimum mean square error criterion (MMSE) filtering to obtain the channel estimation result of each MU-MIMO user.

This step is used to denoise the LS channel estimation sequence of each user, so as to obtain the final channel estimation result of each user. Here, the difference from the existing MMSE method is that the filtering is performed based on the LS channel estimation sequence of a single user, so that the order of the filter is small and easy to implement.

The specific implementation of the MMSE filtering in this step is within the grasp of those skilled in the art, and will not be repeated here.

It can be seen from the above scheme that according to the characteristics of the LTE uplink reference signal, the present invention clarifies that under the condition of minimizing the mutual interference after the MU user pilot superposition, among the N users of MU-MIMO, user n+1 and user n send pilot signals between dealta_n _cs =12/N, thus analyzing the code division multiplexing (CDM) multiplexing relationship of N-order DFT between N users sending pilot signals; Estimation, the single-user channel estimation covers the content of the N-user channel response in MU-MIMO; based on the single-user channel estimation, combined with the characteristics of the N-user transmission pilot signal of MU-MIMO, the LS estimation of the original received signal for user 0 is extracted The relationship between the sequence and the respective channel responses of N users; in the general case where the channel coherence bandwidth is greater than one PRB (ie, 12 subcarriers), the user 0 reference signal is used to estimate the received signal by segmented N-order IDFT, channel separation One subcarrier point for each N-point LS channel estimation of each user is obtained. In this way, the complexity of N-user MU-MIMO channel estimation is reduced, a shortened LS estimation sequence is obtained, and the calculation processing load is also reduced.

In order to evaluate the performance of the algorithm of the present invention, based on the spatial channel model (spatialChannel Model–Extension, SCME) commonly used in LTE systems, linear filters of different groups and orders are used to perform MMSE filtering on the LS channel estimation of each user of 3-user MU-MIMO , the simulation performance curves of the micro-cell SCME model terminal moving speed 3Km/h can be obtained as shown in Figure 2, Figure 3, and Figure 4, and the simulation performance curves of the macro-cell SCME model terminal moving speed 3Km/h are shown in Figure 5, Shown in Figure 6 and Figure 7.

As can be seen from Fig. 2, Fig. 3, and Fig. 4, in the scenario where the micro cell SCME model and the mobile speed of the terminal are 3Km/h, the channel estimation method provided by the present invention adopts 24*8 (24 groups of filters, each The number of filter taps is 8, the following description is similar) MMSE filtering, the performance is better than the 12*12MMSE filtering and 24*24MMSE filtering of the conventional algorithm; and as shown in Figure 5, Figure 6, and Figure 7, in the macro cell SCME model terminal In the scenario where the moving speed is 3Km/h, the channel estimation method provided by the present invention adopts 24*8MMSE filtering, and its performance is better than the 12*12MMSE filtering of the conventional algorithm, and its performance is close to that of 24*24MMSE filtering. It can be seen that the link performance of the channel estimation method proposed by the present invention is better than that of the conventional algorithm, and the performance is close to that of the conventional algorithm whose filter complexity is twice as high.

Under the MU-MIMO application of the conventional LMMSE channel estimation method, its processing flow and complexity are double that of a single user (SU). The channel estimation method provided by the present invention significantly simplifies the MU-MIMO channel estimation of LTE uplink service N users. Computational complexity, and guaranteed user channel estimation performance, provides a better choice for multi-user MU-MIMO channel estimation.

To sum up, the above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (3)

1. A kind of 1. uplink multi-users mimo channel method of estimation, it is characterised in that including：

   A, base station utilizes the up multi-user's multiple-input, multiple-output MU-MIMO user UE for carrying out resource multiplex transmission₀Pilot frequency sequence S₀ (k, l), least square LS channel estimations are carried out to the MU-MIMO reference signal Y (k, l) received, obtain channel estimation resultsWherein, k is carrier index, k=1,2 ..., K；K is the subcarrier number per resource block, and l estimates for the current channel that carries out The numbering of the OFDM frequency pilot signs of meter；

   B, it is described by KIt is divided into K/N sections and carries out N ranks discrete fourier conversion DFT transform, obtains progress resource and answer With the LS channel estimation sequences of each MU-MIMO user of transmission, wherein, N is the MU-MIMO user for carrying out resource multiplex transmission Number；

   C, for the LS channel estimation sequences for each MU-MIMO user for carrying out resource multiplex transmission, carry out respectively minimum Mean-square error criteria MMSE is filtered, and obtains the channel estimation results of each MU-MIMO user.
2. 2. according to the method for claim 1, it is characterised in that the step b includes：

   Using describedAccording toObtain each MU-MIMO The least square LS channel estimation sequences of userWherein, m=0,1, 2 ..., K/N-1, n are Customs Assigned Number.
3. 3. according to the method for claim 1, it is characterised in that in step a according toCarry out The LS channel estimations, wherein, S^* ₀(k, l) is S₀The conjugation of (k, l).